2025 Volume 72 Issue 2 Pages 229-237
The correlation of obesity and metabolic abnormalities with asthma and non-alcoholic hepatic steatosis has been extensively studied. However, the association between asthma and non-alcoholic hepatic steatosis has been largely overlooked. This study aims to investigate the potential association between asthma risk and the fatty liver index (FLI), a validated indicator of non-alcoholic fatty liver disease (NAFLD). We screened 16,223 adults from National Health and Nutrition Examination Survey (NHANES) data between 2001 and 2018. Logistic regression analysis was performed to identify the association between FLI and asthma risk. We assessed their dose-response relationship using a restricted cubic spline (RCS) model. The threshold effect was analyzed to identify the FLI threshold point. Among the subjects screened, there were 2,192 cases suffered from asthma. After adjusting for all the confounders, using the Q3 group (FLI, 54–83) as the reference, the odds ratios (ORs) were 1.35 for the Q1 group (95% CI, 1.01–1.81), 1.21 for Q2 (95% CI, 0.98–1.49), and 1.48 for Q4 (95% CI, 1.27–1.73). Moreover, the RCS showed a nonlinear relationship between FLI and asthma risk (p < 0.05). Although the nonlinear relationship remained significant after gender-based stratification (p < 0.05), low FLI did not confer an increased risk of asthma in females. The optimal FLI threshold was 65 for the study sample; it was 68 and 63 for males and females, respectively (p < 0.05). This study demonstrated a nonlinear relationship between FLI and asthma risk. Furthermore, maintaining respective index values of 68 and 63 for males and females is likely associated with the lowest asthma risk.
Asthma is a chronic pulmonary disease, characterized by persistent airway inflammation, which arises from the infiltration of diverse inflammatory cells, such as eosinophils, neutrophils, lymphocytes, macrophages, and mast cells. Clinically, it presents with recurrent attacks of wheezing, coughing, dyspnea, and chest tightness, with respiratory symptoms exhibiting variable intensity and manifestation over time, commonly worsening during nocturnal or early morning periods. Despite often being categorized as a reversible airway obstruction, asthma has the propensity to evolve into irreversible impairment of pulmonary function if not properly managed [1]. It is also a complex heterogeneous disease with high morbidity, affecting over 300 million individuals worldwide [2]. The traditional causes of asthma are multifactorial, including environmental and host immunological and genetic factors [3]. Furthermore, the importance of metabolic abnormalities in asthma risk has recently been increasingly recognized. Asthma is frequently associated with metabolic abnormalities such as obesity and insulin resistance (IR). In the United States, 60% of patients with severe asthma are obese [4]. Obesity is an independent risk factor for asthma, resulting in poor control and frequent exacerbations [5]. Several studies have identified the distinct clinical features of obesity-related asthma [6, 7]. Litonjua et al. have demonstrated in a male cohort that the development of airway hyperresponsiveness (AHR) was linked to both high and low body mass index (BMI) [8]. Moreover, insulin itself may activate airway immune and structural cells that cause inflammation and narrowing of the airways [9].
Therefore, metabolic factors are significant in the development of asthma. Notably, hepatic steatosis is a primary characteristic of metabolically unhealthy obesity, as it is closely associated with insulin resistance in the liver, muscle, and adipose tissue [10-13]. Several studies revealed a positive correlation between NAFLD and the occurrence of extrahepatic tumors, diabetes, cardiovascular disease, and metabolic syndrome [14]. However, hepatic steatosis has rarely been mentioned in the development of asthma. Asthma risk among adults was possibly influenced by NAFLD as measured by fatty liver index (FLI) [15]. FLI is a widely used index for predicting fatty liver in the world, including triglycerides (TG), BMI, γ-glutamyltransferase and waist circumference. FLI is an accurate and easy indicator as TG, BMI, γ-glutamyltransferase and waist circumference are readily accessible from clinical data [16]. While liver biopsy is considered the gold standard for diagnosing NAFLD, its invasiveness, the likelihood of sampling error, and high complication rate render it unnecessary for all NAFLD patients. Therefore, non-invasive, non-imaging methods have been explored for diagnosing fatty liver, such as the FLI, hepatic steatosis index and the NAFLD liver fat score. Apart from extensive studies on FLI-based surrogates for hepatic steatosis, recent studies have revealed that the degree of hepatic steatosis exacerbates with increasing FLI [17, 18].
A large population-based study with a median follow-up period of 8.3 years revealed a linear increase in mortality risk for all-cause and disease-specific causes, such as cardiovascular disease (CVD), cancer, respiratory disease, and liver disease, with increasing FLI scores [19]. However, the relationship between FLI and asthma has been scarcely reported. Since both high and low BMI are proven to be associated with AHR, we hypothesize a nonlinear relationship between FLI and asthma [8]. Therefore, we attempt to clarify the relationship between FLI and asthma risk and explore its nonlinearity.
The data for this study were obtained from nine cycles of the National Health and Nutrition Examination Survey (NHANES) conducted between 2001 and 2018. NHANES employed a cross-sectional design to investigate demographic, socio-economic, health, and nutritional data at the national level in the United States. All study participants provided informed consent. The NHANES investigation has obtained approval from the National Center for Health Statistics Research Ethics Review Board. Furthermore, explicit informed consent was deemed unnecessary as the data was analyzed secondarily to public data. The present report adheres to the guidelines outlined in Strengthening the Reporting of Observational Studies in Epidemiology.
Inclusion criteria: 1) Subjects ≥20 years old; 2) Available fasting blood collection data. Exclusion criteria: 1) Subjects lacking alcohol consumption data and those excessive alcohol drinkers; 2) Subjects infected with hepatitis B and C viruses and those with liver cancer; 3) Subjects who lack asthma data; 4) Subjects with incomplete data.
Of the 21,634 participants identified by the inclusion criteria, participants lacking alcohol consumption data (n = 1,230), excessive alcohol drinkers (n = 2,172), and those with other types of chronic liver disease such as viral hepatitis (confirmed by HCV-RNA, HCV-antibody, or hepatitis B surface antigen[HBsAg]) and liver cancer (n = 99) were excluded. All the included participants had asthma-related data.
Participants who had missing data on variables of age, gender, race, educational level, smoking status, alcohol consumption, diabetes, family history, BMI, high-density lipoprotein (HDL)-cholesterol, TG and γ-glutamyltransferase were excluded (n = 1,910). Finally, 16,223 adults were selected as our subjects.
VariablesWe used known confounders from clinical practice and previous methods for determining covariates. We included the categorical variables of gender, race, educational level, smoking status, alcohol drinking, family history, BMI, and diabetes, and continuous variables of age and HDL-cholesterol as covariates.
Asthma was identified as a positive reply to the question, “Has a doctor or other health professional ever diagnosed you with asthma?” Patients were categorized based on race as Hispanic, non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, and others, and on their educational level as less than high school, high school/equivalent, and college or higher. Individuals who answered negatively to the inquiry regarding whether they had ever smoked 100 cigarettes throughout their lifetime were regarded as ‘never smokers.’ Subsequently, individuals who responded positively were asked if they currently engage in the activity to distinguish between ‘current smoker’ and ‘former smoker.’ Participants were categorized as ‘current smokers’ if they responded affirmatively to the question and as ‘former smokers’ if they had previously smoked at least 100 cigarettes and were not currently smoking. Participants who reported consuming more than twelve alcoholic drinks annually were classified as ‘alcohol drinker.’ Family history was determined by a positive response to the question, “Blood relatives have asthma.” Furthermore, patients were classified based on BMI as underweight (<18.5 kg/m2), normal-weight (18.5–24.9 kg/m2), overweight (25.0–29.9 kg/m2), or obese (≥30.0 kg/m2). Diabetic participants were identified using the following criteria: (a) fasting plasma glucose ≥126 mg/dL or hemoglobin A1C ≥6.5% [19]; (b) positive response to the question, “Has a doctor confirmed that you have diabetes?” A 24-hour recall was conducted to examine daily alcohol consumption. Excessive alcohol drinking was defined as consuming >30 grams per day for men and >20 grams per day for women [20]. The average alcohol consumption was calculated for participants who completed both 24-hour recalls, while the first recall was used for others.
Bedogni et al. calculated FLI by the formula [15]: FLI = (e0.953 × loge (triglycerides, mg/dL) + 0.139 × BMI, kg/m2 + 0.718 × loge (γ-glutamyltransferase, U/L) + 0.053 × waist circumference, cm – 15.745))/(1 + e0.953 × loge (triglycerides, mg/dL) + 0.139 × BMI, kg/m2 + 0.718 × loge (γ-glutamyltransferase, U/L) + 0.053 × waist circumference, cm – 15.745) × 100.
Statistical analysisStatistical analysis was performed using R software (version 4.1.0, R Foundation for Statistical Computing, Vienna, Austria) and EmpowerStats statistical software (http://www.empowerstats.com). Continuous variables with normal distribution and without normal distribution were compared by ANOVA and Kruskal–Wallis test, respectively. Categorical variables were reported as frequency and percentage and compared using the chi-square test. The subjects were categorized into quartile groups based on their FLI scores (Q1, 0–22; Q2, 23–53; Q3, 54–83; Q4, 84–100).
We adjusted for age in model 1 using univariate logistic analysis. Furthermore, we adjusted for age, gender, race, educational level, alcohol drinking, smoking status, and family history using multivariate logistic analysis in model 2; we adjusted for all confounders, including age, gender, race, educational level, alcohol drinking, smoking status, family history, BMI, diabetes, and HDL-cholesterol into the logistic regression in model 3. The Q3 group served as the reference group as it had the lowest risk of asthma. Using the restricted cubic spline (RCS) in fully adjusted model 3, we evaluated the nonlinear association between FLI and asthma risk. Moreover, we demonstrated the nonlinear relationship between asthma risk and FLI for men and women through gender-based stratification analysis. The RCS model evaluated the association between FLI and asthma risk using four knots (5th, 35th, 65th, and 95th percentiles of FLI).
Finally, we obtained optimal inflection points of the curves by analyzing threshold effects using smooth curve fitting. We established multiple models by treating specific percentiles (between 10% and 90%) of FLI as inflection points and performed piecewise linear regression using them as inputs. Using the log-likelihood ratio test, we compared the one-line (non-segmented) model with the piecewise (segmented) regression model to find the optimal inflection point. Statistical significance was set at p < 0.05.
Among the 16,223 participants, 7,668 were men, and 8,555 were women. The average participant age was 47.24 ± 0.25 years. Participants with asthma were included in the case group (n = 2,192), and the others (n = 14,031) were considered as controls. The subjects were further subdivided into four quartile groups based on their FLI scores, namely Q1 (n = 4,013), Q2 (n = 3,976), Q3 (n = 4,103), and Q4 (n = 4,131). The FLI quartile groups were compared for differences in each component; the groups showed significant differences in age, gender, race, educational level, smoking status, family history, BMI, diabetes, HDL-cholesterol, TG and γ-glutamyltransferase (p < 0.05) (Table 1).
| Characteristics | Total n = 16,223 |
Q1 n = 4,013 |
Q2 n = 3,976 |
Q3 n = 4,103 |
Q4 n = 4,131 |
p-value |
|---|---|---|---|---|---|---|
| Age (years) | 47.24 ± 0.25 | 41.41 ± 17.22 | 49.28 ± 17.37 | 50.07 ± 16.17 | 48.96 ± 15.40 | <0.001 |
| Gender (%) | <0.001 | |||||
| Male | 7,668 | 1,473 (36.71%) | 1,939 (48.77%) | 2,181 (53.16%) | 2,075 (50.23%) | |
| Female | 8,555 | 2,540 (63.29%) | 2,037 (51.23%) | 1,922 (46.84%) | 2,056 (49.77%) | |
| Race (%) | <0.001 | |||||
| Hispanic | 2,880 | 493 (12.29%) | 705 (17.73%) | 883 (21.52%) | 799 (19.34%) | |
| non-Hispanic White | 1,437 | 314 (7.82%) | 397 (9.98%) | 378 (9.21%) | 348 (8.42%) | |
| non-Hispanic Black | 7,322 | 1,873 (46.67%) | 1,709 (42.98%) | 1,816 (44.26%) | 1,924 (46.57%) | |
| non-Hispanic Asian | 3,171 | 767 (19.11%) | 774 (19.47%) | 754 (18.38%) | 876 (21.21%) | |
| other races | 1,413 | 566 (14.10%) | 391 (9.83%) | 272 (6.63%) | 184 (4.45%) | |
| Educational level (%) | <0.001 | |||||
| less than high school | 4,189 | 773 (19.26%) | 1,065 (26.79%) | 1,163 (28.35%) | 1,188 (28.76%) | |
| high school/equivalent | 3,768 | 864 (21.53%) | 897 (22.56%) | 1,049 (25.57%) | 958 (23.19%) | |
| college or above | 8,266 | 2,376 (59.21%) | 2,014 (50.65%) | 1,891 (46.09%) | 1,985 (48.05%) | |
| BMI (%) | <0.001 | |||||
| underweight | 240 | 232 (5.78%) | 8 (0.20%) | 0 (0.00%) | 0 (0.00%) | |
| normal-weight | 4,363 | 3,119 (77.72%) | 1,083 (27.24%) | 154 (3.75%) | 7 (0.17%) | |
| overweight | 5,493 | 647 (16.12%) | 2,483 (62.45%) | 2,049 (49.94%) | 314 (7.60%) | |
| obese | 6,127 | 15 (0.37%) | 402 (10.11%) | 1,900 (46.31%) | 3,810 (92.23%) | |
| Diabetes (%) | <0.001 | |||||
| No | 13,274 | 3,802 (95.12%) | 3,448 (86.81%) | 3,260 (79.55%) | 2,764 (66.50%) | |
| Yes | 2,949 | 195 (4.88%) | 524 (13.19%) | 838 (20.45%) | 1,392 (33.50%) | |
| Family history (%) | <0.001 | |||||
| No | 12,840 | 3,134 (78.10%) | 3,216 (80.89%) | 3,304 (80.53%) | 3,186 (77.12%) | |
| Yes | 3,383 | 879 (21.90%) | 760 (19.11%) | 799 (19.47%) | 945 (22.88%) | |
| Smoking (%) | <0.001 | |||||
| Never | 9,078 | 2,496 (62.20%) | 2,260 (56.84%) | 2,239 (54.57%) | 2,083 (50.42%) | |
| Ever | 4,669 | 827 (20.61%) | 1,160 (29.18%) | 1,282 (31.25%) | 1,400 (33.89%) | |
| Current | 2,476 | 690 (17.19%) | 556 (13.98%) | 582 (14.18%) | 648 (15.69%) | |
| Alcohol drinker (%) | <0.292 | |||||
| No | 3,627 | 886 (22.08%) | 908 (22.84%) | 946 (23.06%) | 887 (21.47%) | |
| Yes | 12,596 | 3,127 (77.92%) | 3,068 (77.16%) | 3,157 (76.94%) | 3,244 (78.53%) | |
| HDL-cholesterol (mmol/L) | 1.37 ± 0.01 | 1.62 ± 0.42 | 1.41 ± 0.37 | 1.28 ± 0.33 | 1.16 ± 0.30 | <0.001 |
| TG (mg/dL) | 107.09 (10.00–4,229.50) |
69.92 (10.00–323.03) |
100.89 (17.97–533.57) |
128.86 (17.97–856.33) |
153.90 (20.00–4,229.50) |
<0.001 |
| γ-glutamyltransferase (IU/L) | 20.00 (4.00–1,681.00) |
14.00 (4.00–286.00) |
18.00 (4.00–223.00) |
22.00 (5.00–659.00) |
28.00 (5.00–1,681.00) |
<0.001 |
Q1, 0 < FLI ≤ 22; Q2, 22 < FLI ≤ 53; Q3, 53 < FLI ≤ 83; Q4, 83 < FLI ≤ 100; FLI, fatty liver index; BMI, body mass index; HDL, high-density lipoprotein; TG, triglycerides.
In age-adjusted model 1, the ORs were 1.01 (95% CI, 0.83–1.22) for Q1, 0.92 for Q2 (95% CI, 0.77–1.10), and 1.57 for Q4 (95% CI, 1.33–1.86). Based on model 1, gender, race, educational level, smoking, alcohol drinking, and family history were incorporated into the logistic regression in model 2; the ORs were 0.98 (95% CI, 0.82–1.18) for Q1, 1.06 for Q2 (95% CI, 0.88–1.28), and 1.64 for Q4 group (95% CI, 1.40–1.92). When adjusted for BMI, diabetes, and HDL-cholesterol based on model 3, the ORs were 1.35 (95% CI, 1.01–1.81) for Q1, 1.21 (95% CI, 0.98–1.49) for Q2, and 1.48 for Q4 (95% CI, 1.27–1.73). In the fully adjusted model 3, as the FLI scores increased, the asthma risk decreased initially, followed by an increase to the highest OR in the Q4 group (Table 2).
| Model 1 | p-value | Model 2 | p-value | Model 3 | p-value | ||||
|---|---|---|---|---|---|---|---|---|---|
| OR | 95%CI | OR | 95%CI | OR | 95%CI | ||||
| Q1 | 1.01 | 0.83, 1.22 | 0.944 | 0.98 | 0.82, 1.18 | 0.835 | 1.35 | 1.01, 1.81 | 0.043 |
| Q2 | 0.92 | 0.77, 1.10 | 0.349 | 1.06 | 0.88, 1.28 | 0.512 | 1.21 | 0.98, 1.49 | 0.079 |
| Q3 | reference | reference | reference | ||||||
| Q4 | 1.57 | 1.33, 1.86 | <0.001 | 1.64 | 1.40, 1.92 | <0.001 | 1.48 | 1.27, 1.73 | <0.001 |
Model 1, we adjusted age in the univariate logistic analysis; Model 2, age, gender, race, educational level, alcohol drinking, smoking status and family history were adjusted. Model 3, age, gender, race, educational level, alcohol drinking, smoking status, family history, BMI, diabetes and HDL-cholesterol were adjusted. Q1, 0 < fatty liver index ≤ 22; Q2, 22 < fatty liver index ≤ 53; Q3, 53 < fatty liver index ≤ 83; Q4, 83 < fatty liver index ≤ 100.
Furthermore, we utilized the RCS based on model 3 to explore the dose-response relationship between FLI and the risk of asthma. The findings indicated a nonlinear relationship between FLI and asthma risk (p < 0.05) (Fig. 1). The figure illustrates that the risk of asthma gradually decreased as FLI scores increased from the Q1 to Q3 quartile values and significantly increased henceforth. In subgroup analysis, we observed a nonlinear relationship in both men and women (p < 0.05) (Figs. 2, 3). Among men, the risk of asthma also decreased gradually with increasing FLI scores from the Q1 to Q3 quartile values, followed by a noticeable increase. High and low FLI scores were significantly associated with a high risk of asthma. Among women, the risk of asthma did not decrease with the increasing FLI scores initially and then established an upward trend from the Q3 quartile value.
Based on model 3, adjusted for age, gender, race, educational level, alcohol drinking, smoking status, family history, BMI, diabetes, and HDL-cholesterol, the threshold analysis demonstrated the turning points of the curves at 65, 68, and 63 (p < 0.05), signifying the FLI values corresponding to the lowest asthma risk for all participants, men, and women, respectively (Table 3).
| Total population | p-value | Males | p-value | Females | p-value | |
|---|---|---|---|---|---|---|
| (n = 16,233) | (n = 7,668) | (n = 8,555) | ||||
| Turingpoint | 65 | 68 | 63 | |||
| <Turingpoint | 0.996 (0.992, 1.000) | 0.054 | 0.991 (0.985, 0.997) | 0.006 | 0.998 (0.993, 1.004) | 0.604 |
| >Turingpoint | 1.017 (1.011, 1.023) | <0.001 | 1.015 (1.004, 1.025) | 0.006 | 1.020 (1.013, 1.027) | <0.001 |
| p for log likelihood ratio test | <0.001 | <0.001 | <0.001 |
In recent years, metabolic risk factors for asthma have received increased attention due to their significant role in the disease and ease of monitoring and control. The effect of obesity on asthma has been a great concern as it is related to the severity of the condition [5]. Individuals with morbid obesity frequently exhibit a modified adipokine secretion profile from malfunctioning adipose tissue that leads to systemic, subclinical inflammation, which hinders the immune response of the lungs and encourages the development of AHR, potentially contributing to the onset of airway inflammation and increased asthma susceptibility [6, 21]. Moreover, IR is prevalent in individuals with asthma and has been associated with decreased pulmonary function, accelerated decline in lung function, and inadequate responses to bronchodilator and corticosteroid therapies. Hepatic steatosis, often associated with metabolically unhealthy obesity and insulin resistance, could be related to asthma. While liver biopsies are the gold standard for assessing hepatic steatosis, they are impractical and invasive for the general population. FLI has been established as a surrogate indicator for hepatic steatosis and has shown a correlation with the degree of hepatic steatosis in previous research [17]. Based on this background, we initiated a population-based study using data from the NHANES database, screening 16,223 individuals over nine periods from 2001 to 2018 to further investigate the relationship between FLI and asthma.
Our study revealed that, after adjusting for age, gender, race, educational level, alcohol drinking, smoking status, family history, BMI, diabetes, and HDL-cholesterol, the risk of asthma exhibited a gradual decrease from group Q1 (OR, 1.35; 95% CI, 1.01–1.81), and was lowest in group Q3 (reference), followed by a significant increase in group Q4 (OR, 1.48; 95% CI, 1.27–1.73). A significant nonlinear relationship was revealed between FLI and asthma by the RCS model used in a fully adjusted logistic analysis (p < 0.05). The curve showed a gradual decline in the risk of asthma, followed by an upward trend approximately from the Q3 quartile. Gender-specific stratified analysis of the nonlinear relationship showed a more pronounced two-end effect in men. Contrastingly, in women, the risk of asthma stabilized initially and started to increase from the Q3 quartile. A previous study indicated a positive association between NAFLD identified by FLI and asthma risk compared to non-fatty liver [15]. However, unlike our study, this research did not further explore the nonlinear relationship between FLI and asthma using RCS or rule out heavy drinkers. Our result revealed a nonlinear relationship between FLI and asthma risk in all subjects. Moreover, we innovatively identified the inflection point of the curve at 65. Fig. 1 shows that the risk of developing asthma decreases gradually initially and increases at an approximate FLI of 65. FLI ≥60 also corresponds to the ultrasonographic detection of fatty liver [16], which is close to our threshold of 65. Hence, maintaining an FLI of approximately 65 leads to the lowest risk of asthma. Additionally, we stratified the results by gender and identified the optimal threshold of 68 for men and 63 for women.
To the best of our knowledge, no prior research has investigated the nonlinear relationship between FLI and the risk of asthma. However, some studies provide partial support for the association. One study focused on precisely measuring adiposity through body fat percentage (BF%) using the bioimpedance technique and found a U-shaped association between BF% and asthma risk in children [22].
The specific mechanism linking hepatic steatosis or FLI with asthma risk has not been studied extensively. However, a previous study showed a significant decrease in forced expiratory volume in the first second and forced vital capacity in adults with NAFLD in Asian and United States populations [23], suggesting a potential relationship between liver steatosis and respiratory function. Furthermore, hepatic steatosis is a characteristic feature of metabolically unhealthy obesity and IR in the liver, muscle, and adipose tissue [10]. In morbid obesity, an altered adipokine secretion profile from dysfunctional adipose tissue can lead to systemic, low-grade inflammation, which impairs pulmonary immune response and promotes AHR. Moreover, obesity demonstrates a distinct subclinical chronic inflammation that blood analysis can be partially identified [24]. Experimental and clinical evidence suggests that this inflammation may contribute to airway inflammation, reduced lung function, AHR, and asthma exacerbation [25, 26]. In addition to causing liver steatosis, obesity, particularly of the viscera, has been closely associated with asthma [7, 21]. Visceral fat (VFA) releases free fatty acids that can be transported to the liver, potentially contributing to hepatic steatosis and non-alcoholic steatohepatitis. The accumulation of visceral fat leads to increased triglyceride hydrolysis and subsequent delivery of free fatty acids to the liver, which is closely linked to the development of insulin resistance [27]. Researchers have also identified IR as an aggravating factor in asthma control and airway remodeling [28]. Calco et al. have shown that insulin increases the density of sensory nerves and induces bronchoconstriction in obese mice [29]. Moreover, Rautureau et al. found that mice with dietary obesity exhibit lipid deposition and metabolic shifts in the lungs, which share similarities with those observed in the liver [30].
Nowadays, significant emphasis has been placed in elucidating the association between obesity and asthma. It is widely acknowledged that obesity is a contributory factor to the heightened prevalence of asthma and the augmented risk of associated adverse events [31]. However, the interconnection between low BMI, compromised nutritional status, and asthma has received limited concern, marking a novel aspect of our investigation. FLI serves as an indicator of both low BMI and suboptimal nutritional status. In the United States, the Tennessee Department of Health identified Shelby County, which encompasses Memphis, as experiencing the highest incidence of childhood asthma. Notably, certain cities within Shelby County exhibit some of the nation’s highest indices of economic hardship, which is inversely related to low socioeconomic status (SES). Low SES is linked to malnutrition and chronic conditions, including asthma [32]. Research conducted on children with asthma in Memphis, Tennessee, revealed that a significant proportion suffer from vitamin A deficiency, suggesting that enhanced vitamin A intake may play a pivotal role in disease prevention [33]. Similarly, a Korean study indicated a marked increase in asthma prevalence among adults from food-insecure households [34]. Furthermore, low nutritional status often signifies dysbiosis of the gut microbiota and impaired digestion and absorption capabilities. Malnutrition is frequently observed in the elderly, and is often accompanied by gut microbiota dysbiosis, which is associated with an increased susceptibility to chronic respiratory diseases like asthma [35]. Other studies have indicated that patients with asthma possibly lose weight due to the regulation of metabolism and energy consumption [36], which could be associated with the reduction of hepatic steatosis.
Upon examining the data stratified by gender, it is evident from the graphical representation that there is a notable correlation between a lower FLI and an increased risk of asthma in both the general population and among males. This association, however, does not hold within the female subset. The discovery of this distinctive subgroup in females was a novel aspect of our research. Estrogen, recognized as an anti-inflammatory agent, effectively mitigates NLRP3 inflammasome-mediated inflammatory diseases, thereby endowing females with a protective effect against asthma [37]. In postmenopausal females, the decline in estrogen levels may result in the loss of this protective effect, potentially increasing the incidence of asthma. Shah and colleagues conducted a 17-year cohort study involving 353,173 females aged between 46 and 70, which found that there was an association between previous estrogen replacement therapy and a reduced risk of asthma exacerbations, with a dose-response relationship observed between the duration of estrogen replacement therapy and the risk of asthma attacks [38].
Moreover, estrogen is conducive to maintaining insulin sensitivity and exerting a hepatoprotective effect. The study on ovariectomized animal models has demonstrated a causal relationship between estrogen deficiency and an increased susceptibility to steatosis [39]. Estrogen binds to estrogen-related receptor α(ERRα), which is essential for the regulation of very low-density lipoprotein synthesis and secretion, facilitating the transport of triglycerides synthesized in the liver to extrahepatic tissues. Estrogen deficiency leads to a reduction in the expression of ERRα in hepatocytes, potentially promoting the expression of lipogenic genes or inhibiting the expression of genes related to fatty acid oxidation, disrupting lipid metabolism and resulting in excessive lipid accumulation in the liver [40]. Additionally, estrogen can influence the distribution of fat in females, reducing visceral fat content, and enhancing insulin sensitivity [41, 42]. Consequently, estrogen deficiency in postmenopausal females leads to a redistribution of fat, with a predominant accumulation in visceral fat, contributing to an increased incidence of central obesity and IR, making them a high-risk group for NAFLD. We speculate that, for individuals with high FLI, apart from the risk factor of fatty liver, estrogen levels might also be relatively low, thereby increasing their risk of asthma. For those with low FLI, despite sharing the same risk factors for asthma, such as malnutrition, low BMI and compromised immunity, as males, their estrogen levels might be relatively high, potentially mitigating the detrimental effects of low FLI on asthma. Hence, unlike males, low FLI does not confer an increased risk of asthma in females.
Our investigation represents the inaugural attempt to evaluate the nonlinear dose-response correlation between FLI and susceptibility to asthma. A nonlinear relationship was demonstrated between FLI and asthma risk, with a threshold of 65 (68 for males and 63 for females). This finding provides a novel perspective on the relationship between hepatic steatosis and asthma and offers new insight into the pathogenesis of asthma. However, there are some limitations in this study. Firstly, the relationship between hepatic steatosis and asthma requires substantiating magnetic resonance imaging findings of the liver or its pathology. Secondly, since our study was cross-sectional, it could not establish a causal relationship between hepatic steatosis and the risk of asthma. Therefore, large-scale prospective studies are needed to elucidate this causal relationship further.
To the best of our knowledge, this is the first study to elucidate the nonlinear relationship between FLI and asthma risk, revealing a distinct overall inflection point of 65 (68 for males and 63 for females). Furthermore, our study uncovered a distinct gender disparity, with females demonstrating no discernible association between low FLI and an increased vulnerability to asthma. These insights imply that hepatic steatosis elicits a biphasic influence on asthma susceptibility, highlighting the necessity for stringent regulation of FLI, particularly in males, to mitigate asthma risk. Such an approach may have potential preventive implications for the onset and progression of asthma. However, further validation of these conclusions through more extensive prospective studies is warranted, along with additional laboratory investigations to interpret the underlying mechanisms.
Not applicable.
We would like to thank the participants and staff of the NHANES project for their efforts.
All authors contributed to the conception and design of the study; Hua Qiao was responsible for interpreting the results of the statistical analysis, criticizing the statistical analysis, and revising the manuscript; Tengfei Sun was responsible for drafting the manuscript, analyzing and interpreting the data; Kexin Fan was responsible for collecting the data and statistical analysis; Zhuoxiao Han contributed in criticism of the design and statistical analysis. The final manuscript was read and approved by all authors.
This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.
The authors declare that they have no conflict of interest.
All data generated or analyzed during this study are available from the corresponding author upon reasonable request.
